W3C home > Mailing lists > Public > public-webapi@w3.org > April 2008

Re: [selectors-api] Handling :link and :visited Pseudo Classes

From: Arve Bersvendsen <arveb@opera.com>
Date: Thu, 17 Apr 2008 14:14:32 +0200
To: "L. David Baron" <dbaron@dbaron.org>, "Maciej Stachowiak" <mjs@apple.com>
Cc: "Travis Leithead" <travil@windows.microsoft.com>, "Lachlan Hunt" <lachlan.hunt@lachy.id.au>, public-webapi <public-webapi@w3.org>
Message-ID: <op.t9q8aihnbyn2jm@galactica>

On Thu, 17 Apr 2008 00:28:36 +0200, L. David Baron <dbaron@dbaron.org>  

> My current thinking (from
> https://bugzilla.mozilla.org/show_bug.cgi?id=147777#c65 ) is that
> what we'd need to do to fix this is:

I'm not sure I "get" this proposal. Are you proposing that the Selectors  
API lie about what goes in to the collection, and that other associated  
APIs simply lie about what the values of any particular property apply? Or  
are you suggesting that the CSS WG go back and alter the specification so  
that only the following properties are allowed on the :visited selector?


I believe all other properties may affect either layout or performance.  
 From the value types: Different color spaces, named colors and alpha  
values may, depending on the implementation, be suspect to timing attacks.

Now, if your proposal is the one in my first question: An implementation  
which chooses to lie about some properties would effectively have to lie  
about all properties, keeping two layouts in memory, given that :visited  
may affect the layout of other elements as well. See Ian's and Maciej note  
on this.

Then again, if I'm to discard the complicator's gloves here: Couldn't UAs  
simply offer a user option to disable support for :visited in stylesheets  

Arve Bersvendsen

Developer, Opera Software ASA, http://www.opera.com/
Received on Thursday, 17 April 2008 12:15:28 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 18:10:01 UTC